Assistive and Autonomous Breast Ultrasound Screening: Improving PPV and Reducing RSI
Abstract
Women with dense breasts are at increased risk for breast cancer. For these women, mammography is less effective in detecting breast cancer. Knowing this, many women, when informed of their breast density, are seeking ultrasound screening in addition to mammography. Ultrasound detects additional breast cancers missed by mammography, but also detects many more lesions that ultimately prove benign when biopsied. The high number of false positives associated with ultrasound screening is costly. In addition to the expense associated with additional biopsies, false positives create unnecessary stress and anxiety. Reducing the rate of false positives is a critical challenge in ultrasound breast screening that would dramatically improve patient care. This work addresses the Breast Cancer Research Program challenge to "Conquer the problems of overdiagnosis and overtreatment." Mechanical properties, such as stiffness, vary between malignant and benign tissues. Furthermore, there is evidence that the mechanical stiffness of tissue is itself a determining factor in the aggressiveness of cancer. In this proposal, we will use a sonographer-directed anthropomorphic robotic arm to enable high-resolution viscoelastic and non-linear elastography, providing images and quantification of the mechanical properties of tissue. We will exploit the precision motion and force-sensing capabilities of a robotic arm in combination with ultrasound imaging to measure the stiffness and viscosity of breast lesions. We expect that this additional information will allow improved discrimination of malignant and benign lesions, and allow a reduction in the number of negative biopsies. While some ultrasound systems currently feature elastography (stiffness-sensing) capabilities, they suffer from operator dependence and lack of repeatability, due in part to variations in transducer placement and pressure. Our hypothesis is that the precision control over transducer positioning enabled by the robotic arm will enable advanced ultrasound elastography techniques that are infeasible with handheld scanning and that these techniques will improve lesion classification, reduce false positives, and reduce unnecessary biopsies. In addition to the reduction in false positives, there are other advantages to this approach. In contrast to current automated breast scanners, our system follows the natural shape of the breast in a manner similar to an expert sonographer. Existing automated systems flatten the breast and drag the probe in a straight line over the breast, and as a result suffer from image artifacts near the chest wall and edges of the field of view. Our proposed system avoids these complications by allowing the sonographer to use the robotic arm to interactively examine the breast, rather than relying on a static image obtained by existing automated systems. Furthermore, the robotic arm takes the load of scanning off the sonographers arm, reducing the potential for repetitive stress injury. Our efforts seek to help patients undergoing breast screening, especially those with mammographically dense breast tissue -- about a third of all women, and half of younger women. While this work will not end breast cancer, it does have the potential to significantly improve clinical care for those undergoing screening by making ultrasound breast imaging more predictive of malignancy. A risk associated with more widespread ultrasound screening is increased overdiagnosis and overtreatment; this work aims to mitigate that risk while providing the benefit of increased early cancer detection. The long-term goal of this research is to increase both the utility and availability of ultrasound elastography in breast screening. We expect that we can achieve a patient-outcome within 5 years of the completion of this study by taking advantage of robotic and autonomous planning technologies already in development for other applications. Clinical trials subseque
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 31, 2017
- Source ID
- W81XWH1710021
Entities
People
- Stephen McAleavey
Organizations
- United States Army
- University of Rochester